Reproducible example:
require(ggplot2)
set.seed(7)
# generate data
a <- runif(100, min=0, max = 50000)
c <- sample(c("A", "B"), 100, replace =T)
c <- factor(c)
df <- data.frame(a,c)
cdplot(c ~ a, data = df)
abline(h=0.5, lwd=2, lty=2, col = "orange")
gg <- ggplot(df, aes(a, fill = c)) + geom_density(position='fill', alpha = 0.5)
gg <- gg + geom_hline(yintercept=0.5, linetype = "dashed", size = 2)
gg
Here is an image of the two plots superimposed:

The two curves are clearly different, e.g. at around x = 45000 one reaches y = 0.5 while the other doesn't. In my real data the differences are way bigger. Does anyone know what the reason for this is, and, well, which one I should believe? In my real data it's a fairly big deal whether the conditional probability for a variable reaches 0.5 or not.